Abstract
BACKGROUND: Lung adenocarcinoma (LUAD) is the most common histological subtype of lung cancer. The distinctive genetic and epigenetic modifications in tumors and paired non-malignant samples, such as adjacent peri-tumor and tumor-distant normal lung tissues, have not been adequately studied. METHODS: We recruited 57 patients with resectable stage I-III LUAD and collected matched samples of the primary tumor, peri-tumoral tissues, and tumor-distant normal lung tissue. We performed bisulfite sequencing using a custom methylation panel to profile DNA methylation levels and obtained somatic variation landscape through targeted next-generation sequencing (NGS). We attempted to identify differential methylation blocks (DMBs) between the tumor, peri-tumor, and normal tissues. RESULTS: We analyzed the DNA methylation patterns of matched tumor, peri-tumor, and normal lung tissue samples from 57 LUAD patients. No significantly different methylation blocks were found between peri-tumoral and normal tissues, while they both exhibited distinct methylation profiles compared to tumor tissues. A total of 1329 tumor-specific DMBs, which are potentially associated with aberrant gene expression in LUAD, were identified. Utilizing a consensus clustering algorithm, we classified the tumor samples into two subgroups (C1 and C2) based on distinct methylation profiles, independent of the patient's sex, tumor stage, smoking history, and tumor cell fraction. The C2 subgroup exhibited a higher malignancy density ratio (MD ratio), suggesting a more pronounced degree of field cancerization, while the C1 subgroup was characterized by a higher frequency of EGFR mutations. The DMBs between the two subgroups were enriched in the calcium signaling pathway. Notably, P2RX2 shows significant hypermethylation in the C2 subgroup, and its low expression in the external The Cancer Genome Atlas (TCGA) cohort may correlate with reduced overall survival in LUAD patients. CONCLUSION: Our findings revealed distinct methylation patterns between tumor and pre-malignant samples, such as peri-tumor and normal tissues. Moreover, our study suggests that distinct clustering based on DNA methylation may indicate different prognoses in LUAD patients.